Home A - Z Create page
Page Index

An Otter Wiki

Blame

75fbd9 Anonymous 2026-04-15 07:58:32 1 
# Indra-Swarm API
6ac36d Anonymous 2026-04-15 07:50:35 2
3
## Gemma-4-26B-A4B-it
3b63b8 Anonymous 2026-04-15 07:52:30 4
7b949e Anonymous 2026-04-15 07:57:05 5
[Model info](https://huggingface.co/google/gemma-4-26B-A4B-it)
6ac36d Anonymous 2026-04-15 07:50:35 6
7
**Docker image's Current Sglang Config**
fc5c99 Anonymous 2026-04-15 07:51:06 8
9
d5a95c Anonymous 2026-04-15 07:52:06 10
```
6ac36d Anonymous 2026-04-15 07:50:35 11
python3 -m sglang.launch_server
12
--model-path google/gemma-4-26b-a4b-it
13
--tp 2
14
--port 3000
15
--host 0.0.0.0
16
--attention-backend triton
17
--mem-fraction-static 0.8
18
--max-running-requests 128
19
--chunked-prefill-size 4096
20
--context-length 32768
21
--trust-remote-code
22
--enable-piecewise-cuda-graph
23
--schedule-policy lpm
d5a95c Anonymous 2026-04-15 07:52:06 24
```
e30422 Anonymous 2026-04-15 07:50:55 25
6ac36d Anonymous 2026-04-15 07:50:35 26
**Test Curl**
e30422 Anonymous 2026-04-15 07:50:55 27
7b949e Anonymous 2026-04-15 07:57:05 28
```
4ccd72 Anonymous 2026-04-15 08:16:19 29
curl http://192.168.40.40:3000/v1/chat/completions \
7b949e Anonymous 2026-04-15 07:57:05 30
-H "Content-Type: application/json" \
31
-d '{
32
"model": "google/gemma-4-26b-a4b-it",
33
"messages": [{"role": "user", "content": "System check. Are you online?"}]
34
}'
35
```
36
37
## faster-Qwen3-tts
38
39
[Model info](https://github.com/andimarafioti/faster-qwen3-tts)
40
41
**Test Curl**
42
d5a95c Anonymous 2026-04-15 07:52:06 43
```
b4a67b Anonymous 2026-04-16 06:00:40 44
curl -X POST http://192.168.40.40:8002/v1/audio/speech \
45
-H "Content-Type: application/json" \
46
-d '{
47
"model": "tts-1",
48
"input": "This is a text-to-speech system check. Audio synthesis is functional on Indra.",
49
"voice": "nona",
50
"response_format": "wav",
6ac36d Anonymous 2026-04-15 07:50:35 51
"seed": 42
b4a67b Anonymous 2026-04-16 06:00:40 52
}' \
53
--output tts_test.wav
d5a95c Anonymous 2026-04-15 07:52:06 54
```
690379 Anonymous 2026-04-15 08:14:25 55
e435db Anonymous 2026-04-16 06:02:18 56
to change voices, set "voice" to any of the following:
57
- aus-female-1
58
- aus-female-2
59
- aus-female-3
60
- aus-female-4
61
- aus-female-5
62
- aus-female-6
63
- aus-male-1
64
- aus-male-2
65
- aus-male-3
66
- aus-male-4
67
- aus-male-5
68
- aus-male-6
69
- aus-male-7
70
- charter
71
- gaius
72
- _gantry
73
- nona
74
- oni
75
- vulcan
690379 Anonymous 2026-04-15 08:14:25 76
77
78
79
## faster-whisper-large-v3-turbo-ct2
80
81
[Model info](https://huggingface.co/deepdml/faster-whisper-large-v3-turbo-ct2)
82
83
**Test Curl**
84
Uses a known locally saved audio file on the indra machine for testing
85
86
```
6ea707 Anonymous 2026-04-15 08:14:58 87
curl http://192.168.40.40:8005/v1/audio/transcriptions \
690379 Anonymous 2026-04-15 08:14:25 88
-H "Content-Type: multipart/form-data" \
89
-F "file=@/mnt/nvme3n1/swarm/voice-samples/aus-male-1.wav" \
90
-F "model=deepdml/faster-whisper-large-v3-turbo-ct2" \
91
-F "response_format=json"
92
```